Modified fuzzy ants clustering approach |
| |
Authors: | Siriporn Supratid Hwajoon Kim |
| |
Affiliation: | (1) School of Computer and Information Technology, Liaoning Normal University, Dalian, 116029, China;(2) State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, 210093, China |
| |
Abstract: | Being trapped in local optima within clustering search space currently is nontrivial difficulty. In order to relieve such
a difficulty, even using genetic algorithm to optimize the initial clusters for fuzzy c-means is still unsatisfied. Since
genetic algorithm intensifies only the current best solution, it will easily gets trapped in local minima. The ant colony
system, dissimilarly to genetic algorithm, recognizes that the solutions near the best solution are also good ones and they
bring about smoothness of solution. This paper proposes a modified fuzzy ant clustering. Such a presented method is a combination
of genetic algorithm, ant colony system and fuzzy c-means. It is employed in creating fuzzy color histogram in image retrieval
application. The performance measurement relates to the percentages of accuracy of image retrieval. Experimental results show
that the proposed approach yields the best results among others with respect to sensitivity and robustness on dealing with
lighting intensity changes, quantization errors, also changes in number of images and in size of color space, even the certain-range
variation of a particular parameter of clustering. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|